Saying (or even just thinking) a word or phrase results from activity in multiple regions of your brain. Of course, we can measure/'map' this activity to some degree; From wikipedia:

EEG measures the brain's electrical activity directly, while other methods record changes in blood flow (e.g., SPECT, fMRI) or metabolic activity (e.g., PET, NIRS), which are indirect markers of brain electrical activity.

The question is, just how precise is this mapping? Is it precise enough to answer the following question? (and if so, has anyone examined it)?

I'm curious if the same word or thought in different languages is produced by the same pattern of brain activity?

  • For example: does "hello" for a native English speaker generate the same signal response as a Spanish speaker saying "hola"?

More generally, I guess the question I'm asking is: are the signals/activity generated by the brain to create language driven more by the phonetics/orthographics/structure of a word or the meaning/connotation of that word/concept?

  • Can we quantify that person x speaking the greeting "hello" is more/less similar to a Spanish speaker saying "hola" than, say, both speakers saying "a" (in English a is an article; in Spanish a is equivalent to the English "to")?

Two notes / caveats:

  • I would assume that the same word in different languages would generate different signals in the same individual (i.e., one who comprehends the meaning of the word in both languages). Though perhaps this isn't true (which could also be very interesting).

  • Of course perhaps sex, gender, age, or even just general variation between individuals (or even variation within an individual) might produce slightly different activity signals for the same task/word.

  • 1
    $\begingroup$ Based on Christiaan's answer to this question, it does sound like it is in fact potentially possible to differentiate these signals. The question remains: has anyone done this involving language?? $\endgroup$ Commented Feb 25, 2016 at 13:31
  • 1
    $\begingroup$ Hi @WYSIWYG - I have been pondering on this question, but it will take me time to answer it, if I can. I think CogSci is definitely a good alternative, because the question deals with imaging of higher brain functions. Up to you. It's a high-quality question. Thanks for pinging me on this. $\endgroup$
    – AliceD
    Commented Feb 26, 2016 at 7:24
  • $\begingroup$ Bump. Asked this question ~9 months ago and still have not received any response. If you can provide some insight, please comment or answer. Thanks! $\endgroup$ Commented Nov 7, 2016 at 14:38
  • $\begingroup$ I would split this into two questions, which you already identified yourself. What if the second question could be answered without any reference to the first? Or what if somebody knows the answer to the first but not the second? $\endgroup$
    – Steven Jeuris
    Commented Aug 11, 2017 at 7:09

1 Answer 1


Here's an article talking about exactly what you are asking: EEG decoding of spoken words in bilingual listeners: from words to language invariant semantic-conceptual representations

I can say that the brain obviously has the ability to invariantly represent different stimuli as being the same abstract concept, not even from stimuli in the same domain (such as auditory to auditory) but even across different domains of knowledge (identifying astract conceptual relationships between visual signals and motor signals)! Whether this can be derived purely from EEG, I doubt it because it lacks the precision necessary (which would require single cell recordings of millions of cells at the same time). I should also note that EEG can't tell us much on how the brain generates the spectral graphs that we record, just what the electrical activity is. We also don't even know what exactly causes the signals. Electrical activity may not even be the only indication of invariance, because other underlying mechanisms may be involved. But the idea of invariance, itself, is the very idea of what you are talking about and this phenomenon is found in lots of ways. In fact, anytime anyone relates two different stimuli as the same concept, it's clear that somehow the brain has an invariant representation of the event. There exists some kind of mechanism in the brain to recognize and store abstract concepts, removed from the concrete signals from which the concepts are learned.

"Mirror" neurons are an example of the domain crossing invariance, and something you can Google. These neurons demonstrate invariance by being similarly activated when an animal witnesses another animal performing an action and when the animal performs the action itself (e.g. eating).

However, your question about bilingual invariant representations and the mirror neurons are simply specific examples of a more general phenomenon. In fact, the ability for humans to have such incredible invariance is a major substrate that makes our intellectual capabilities so distinctly advanced compared to even the latest and greatest artificial intelligence attempts.

In the auditory language vs. language example you provided, the differences in auditory signals are ignored and our perception of the concept being communicated remains invariant. The same thing happens in vision. Even though the objective visual signals received by the brain are different, objects can be conceptually related. Just look at one face, then look at another one. You will still be able to identify both objects as the singular concept, "face". You might say "but the general structure of the face is still there". Yet, if you actually look at firing patterns of the neurons, they are still completely different. It's just as mysterious how the brain does this as it is that we can identify differing audio signals as the same concept. Humans possess a variety of invariant abilities in object recognition, such as scale invariance (different sizes), rotational invariance, translational invariance, skew invariance, color invariance, contrast invariance, and many others.

Just keep in mind that even though we possess these invariant abilities, we also maintain the distinctions in parallel. So the answer to your question is both yes and no. We generate both an invariant representation and a variant representation of whatever the stimulus is. Theoretically, both representation types are present in the EEG. However, EEG is not powerful enough to identify the specifics of how this works in language and whether this is even true.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.